REMATCH
Resource-efficient tunnelling based on real-time characterisation of the excavated material
The aim of the research project is to utilise the large quantities of soil and rock produced during tunnel construction in a way that conserves resources instead of depositing them in landfill sites, for example.
While excavated material from hard rock can be used comparatively easily, for example as aggregate in concrete, this is much more difficult with loose rock. This is because the conditioned earth pulp has a much lower strength compared to natural geology. For recycling in the sense of the circular economy for other construction purposes (e.g. an earth wall or a road substructure), complex soil treatment is therefore required by adding lime, for example. Although tunnelling machines are equipped with various sensors nowadays, conditioning agents are dosed and added manually because the target and actual values of the soil parameters on the conveyor belt cannot yet be compared.
Therefore, REMATCH aims to carry out a real-time characterisation of excavated material based on artificial intelligence (AI) methods. To this end, a model is being trained that creates a link between representative images and reliably determined soil parameters. Relevant tunnelling, conditioning and extraction data from the TBM are incorporated into the model to increase its informative value. Based on all available data, a preliminary classification is made as to whether the excavated material is "utilisable" or "non-utilisable".
REMATCH is being funded from 2021 to 2024 as part of a bilateral cooperation programme "Promotion of Franco-German projects on the topic of artificial intelligence".
Bundesministerium für Bildung und Forschung (BMBF) und die französische Agence Nationale de la Recherche (ANR)
ARCADIS, CETU - Centre d'Etudes des Tunnels, DB Netze AG, Herrenknecht AG, LIRIS – Labor für Bildverarbeitung und Informationssysteme, TH Köln
2021 to 2024
France, Germany
- Research & Development